aiqbench analysis
Research indicates that Large Language Models significantly reduce the diversity of public debate by converging on a narrow set of polished arguments and fixed structural arcs, capturing only a fraction of the unique perspectives found in human discourse. While prompting for diversity increases variation, models often introduce novel but irrelevant points rather than recovering the specific, concrete arguments humans generate. This phenomenon of "argument collapse" suggests that widespread reliance on LLMs for drafting public-facing content could homogenize long-form discourse across both short and extended essays.
Argument Collapse: LLMs Flatten Long-Form Public Debate
arXiv:2606.01736v3 Announce Type: replace Abstract: As LLMs are increasingly used to draft public-facing arguments, they may flatten public debate by repeatedly introducing the same polished, plausible arguments. We study argument collapse, the tendency of essays generated by different LLMs to converge to a smaller set of main arguments, sub-arguments, and paragraph-level structures. We compare 1,039 human responses from 195 New York Times (NYT) debates, 448 human responses from 61 longer-form Boston Review (BR) forums, and 23,384 LLM-generated essays. In the NYT corpus, 65.3% of human main arguments are unique within a debate, compared to 3.4% of LLM main arguments. Asking LLMs to generate diverse answers adds variation, but a typical model recovers only about half of the distinct human main arguments, with much of the added variation falling outside the observed human argument space. Collapse also appears in sub-arguments, where among essays with the same main argument, 41.0% of human sub-arguments are unique versus 9.1% from LLM responses. Qualitatively, LLMs often reuse generalized and hedged sub-arguments, while humans prefer more concrete and topic-specific ones. Structure-wise, LLM-generated essays tend to follow a more fixed arc, often opening with a direct claim and moving quickly toward proposals. The same patterns hold in longer BR essays, suggesting that argument collapse extends beyond short-form responses.
Abstract:As LLMs are increasingly used to draft public-facing arguments, they may flatten public debate by repeatedly introducing the same polished, plausible arguments. We study argument collapse, the tendency of essays generated by different LLMs to converge to a smaller set of main arguments, sub-arguments, and paragraph-level structures. We compare 1,039 human responses from 195 New York Times (NYT) debates, 448 human responses from 61 longer-form Boston Review (BR) forums, and 23,384 LLM-generated essays. In the NYT corpus, 65.3% of human main arguments are unique within a debate, compared to 3.4% of LLM main arguments. Asking LLMs to generate diverse answers adds variation, but a typical model recovers only about half of the distinct human main arguments, with much of the added variation falling outside the observed human argument space. Collapse also appears in sub-arguments, where among essays with the same main argument, 41.0% of human sub-arguments are unique versus 9.1% from LLM responses. Qualitatively, LLMs often reuse generalized and hedged sub-arguments, while humans prefer more concrete and topic-specific ones. Structure-wise, LLM-generated essays tend to follow a more fixed arc, often opening with a direct claim and moving quickly toward proposals. The same patterns hold in longer BR essays, suggesting that argument collapse extends beyond short-form responses.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2606.01736 [cs.CL]
(or arXiv:2606.01736v3 [cs.CL] for this version)
Full text under the publisher's open licence · 221 words
Related coverage
Argument Collapse: LLMs Flatten Long-Form Public Debate
arXiv · Jun 6, 2026
Argument Collapse: LLMs Flatten Long-Form Public Debate
arXiv · Sep 5, 2026
Jev: System One models for Prod, not God — with Diogo Almeida, CEO, TypeSafe AI
Latent Space · Sep 21, 2026
The Continual Constitutional Convention: How language models revise the documents that govern them
AI Alignment Forum · Oct 13, 2026